To compile computer tomographic representations using a CT with a plurality of angularly offset ray sources, it is generally known to combine data from the individual ray sources and thereby carry out the reconstruction, i.e. a convolution of the data with subsequent back projection, for 2D or 3D reconstruction.
For reconstruction from a dataset of a ray cone which does not fully scan the object, it is furthermore known to supplement the peripherally existing sharp boundaries with corresponding extrapolations of data in order to reduce artifacts which have been created by a sharp delimitation of the data. In this regard, reference is made for example to document DE 198 54 917 A1 which describes such an extrapolation of peripheral data, albeit in a CT with a single-tube/single-detector system.
It has been found that both a simple data supplement in a CT system with a plurality of ray cones, respectively from the other ray cone, can lead to artifacts in the reconstruction of computer tomographic representations.